skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Bayesian based design of real-time sensor systems for high-risk indoor contaminants

Abstract

The sudden release of toxic contaminants that reach indoor spaces can be hazardousto building occupants. To respond effectively, the contaminant release must be quicklydetected and characterized to determine unobserved parameters, such as release locationand strength. Characterizing the release requires solving an inverse problem. Designinga robust real-time sensor system that solves the inverse problem is challenging becausethe fate and transport of contaminants is complex, sensor information is limited andimperfect, and real-time estimation is computationally constrained.This dissertation uses a system-level approach, based on a Bayes Monte Carloframework, to develop sensor-system design concepts and methods. I describe threeinvestigations that explore complex relationships among sensors, network architecture,interpretation algorithms, and system performance. The investigations use data obtainedfrom tracer gas experiments conducted in a real building. The influence of individual sensor characteristics on the sensor-system performance for binary-type contaminant sensors is analyzed. Performance tradeoffs among sensor accuracy, threshold level and response time are identified; these attributes could not be inferred without a system-level analysis. For example, more accurate but slower sensors are found to outperform less accurate but faster sensors. Secondly, I investigate how the sensor-system performance can be understood in terms of contaminant transport processes and the model representation that is used to solve themore » inverse problem. The determination of release location and mass are shown to be related to and constrained by transport and mixing time scales. These time scales explain performance differences among different sensor networks. For example, the effect of longer sensor response times is comparably less for releases with longer mixing time scales. The third investigation explores how information fusion from heterogeneous sensors may improve the sensor-system performance and offset the need for more contaminant sensors. Physics- and algorithm-based frameworks are presented for selecting and fusing information from noncontaminant sensors. The frameworks are demonstrated with door-position sensors, which are found to be more useful in natural airflow conditions, but which cannot compensate for poor placement of contaminant sensors. The concepts and empirical findings have the potential to help in the design of sensor systems for more complex building systems. The research has broader relevance to additional environmental monitoring problems, fault detection and diagnostics, and system design.« less

Authors:
 [1]
  1. Univ. of California, Berkeley, CA (United States)
Publication Date:
Research Org.:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
926499
Report Number(s):
LBNL-125E
DOE Contract Number:
AC02-05CH11231
Resource Type:
Thesis/Dissertation
Resource Relation:
Related Information: Designation of Academic Dissertation: Doctoral; Academic Degree: Doctor of Philosophy in Engineering ? Mechanical Engineering; Name of Academic Institution: University of California, Berkeley; Location of Academic Institution: Berkeley, CA
Country of Publication:
United States
Language:
English
Subject:
54 ENVIRONMENTAL SCIENCES

Citation Formats

Sreedharan, Priya. Bayesian based design of real-time sensor systems for high-risk indoor contaminants. United States: N. p., 2007. Web. doi:10.2172/926499.
Sreedharan, Priya. Bayesian based design of real-time sensor systems for high-risk indoor contaminants. United States. doi:10.2172/926499.
Sreedharan, Priya. Mon . "Bayesian based design of real-time sensor systems for high-risk indoor contaminants". United States. doi:10.2172/926499. https://www.osti.gov/servlets/purl/926499.
@article{osti_926499,
title = {Bayesian based design of real-time sensor systems for high-risk indoor contaminants},
author = {Sreedharan, Priya},
abstractNote = {The sudden release of toxic contaminants that reach indoor spaces can be hazardousto building occupants. To respond effectively, the contaminant release must be quicklydetected and characterized to determine unobserved parameters, such as release locationand strength. Characterizing the release requires solving an inverse problem. Designinga robust real-time sensor system that solves the inverse problem is challenging becausethe fate and transport of contaminants is complex, sensor information is limited andimperfect, and real-time estimation is computationally constrained.This dissertation uses a system-level approach, based on a Bayes Monte Carloframework, to develop sensor-system design concepts and methods. I describe threeinvestigations that explore complex relationships among sensors, network architecture,interpretation algorithms, and system performance. The investigations use data obtainedfrom tracer gas experiments conducted in a real building. The influence of individual sensor characteristics on the sensor-system performance for binary-type contaminant sensors is analyzed. Performance tradeoffs among sensor accuracy, threshold level and response time are identified; these attributes could not be inferred without a system-level analysis. For example, more accurate but slower sensors are found to outperform less accurate but faster sensors. Secondly, I investigate how the sensor-system performance can be understood in terms of contaminant transport processes and the model representation that is used to solve the inverse problem. The determination of release location and mass are shown to be related to and constrained by transport and mixing time scales. These time scales explain performance differences among different sensor networks. For example, the effect of longer sensor response times is comparably less for releases with longer mixing time scales. The third investigation explores how information fusion from heterogeneous sensors may improve the sensor-system performance and offset the need for more contaminant sensors. Physics- and algorithm-based frameworks are presented for selecting and fusing information from noncontaminant sensors. The frameworks are demonstrated with door-position sensors, which are found to be more useful in natural airflow conditions, but which cannot compensate for poor placement of contaminant sensors. The concepts and empirical findings have the potential to help in the design of sensor systems for more complex building systems. The research has broader relevance to additional environmental monitoring problems, fault detection and diagnostics, and system design.},
doi = {10.2172/926499},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Jan 01 00:00:00 EST 2007},
month = {Mon Jan 01 00:00:00 EST 2007}
}

Thesis/Dissertation:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this thesis or dissertation.

Save / Share:
  • This study presents a set of mechanisms for use in the construction of distributed knowledge-based systems that must meet such practical performance constraints. These mechanisms are made manifest in the form of programming-language constructs designed to support the development of DKBS's that implement a multiplicity of alternative policies and abstract architectures. The proposed constructs are studied, using two techniques: implementation in an experimental testbed environment and performance measurement in controlled experiments that assess the time cost of interagent communications among loosely-coupled LISP processes. The author's empirical data on interagent communications should prove valuable to DKBS implementors who must meet real-timemore » constraints using contemporary hardware and software technology. It is concluded that it is possible to construct interagent communications cost models that are useful and have a high degree of predictive value. Experimental results indicate that the cost of message-based interagent communications between symbolic processing agents is very high.« less
  • The objective was the basis of a multiprocessor designed methodology for the implementation of real-time DSP systems represented by recursive shift-invariant flow graphs. This research generalizes and formalizes many aspects of periodic scheduling. It also treats practical considerations associated with implementation of multiple processors in VLSI. Finally, algorithms for the efficient scheduling and mapping of flow graphs are developed. A number of previously unknown fundamentals are revealed, making it possible to solve the design methodology problem within a limited scope. Most important is the unification of periodic scheduling theory, a comprehensive relationship among all periodic MIMD computer architecture classes. Thesemore » relationships allow the transformation of solutions from one class to another. Average fractional processing rates are achieved, using a deterministic and efficient solution method that does not require the problem flow graph to be pre-blocked. Pre-blocking is undesirable because it increases both solution algorithm complexity and final implementation program complexity. This solution method may be used as well for matching the bandwidth of an implementation to the sampling rate of a real-time DSP system. The practical use of ideal predictors along with ideal delays for describing communications requirements is developed. The extra degree of freedom thus provided is intergral to the effective application of generalized (P)SSIMD theory. Efficient flow graph analysis algorithms are also developed. Furthermore, a polynomial time algorithm is developed for scheduling certain classes of problems.« less
  • An analytic model is introduced to estimate task response time for loosely coupled distributed processing systems with real-time applications. The model considers such factors as assignment of modules to computers, module precedence relationships, interprocessor communications, interconnection network delay, and module scheduling policy. Simulation experiments are used to validate the model assumptions and to show the accuracy of the model. The analytic model is first employed to investigate the effects of module precedence relationships on response times. The study reveals that the distributions of module execution times and the mean execution time ratio for a pair of consecutive modules are majormore » factors for the effects. The task response time model is then used to study module assignment for distributed systems. Based on the model, a new local search algorithm for module assignment is developed. The algorithm is validated by applying to two distinct distributed systems for space defense applications.« less
  • The throughput in real-time digital signal processing applications is limited by both the capability of the processors employed for number-crunching operations and the capacity of a supporting communications link. The systolic architectures eliminate the memory bandwidth problems by allowing multiple computations for each memory access and result in a speed-up in the execution time of compute-bound computations. However, the throughput rate in a systolic array is still limited by the computational time needed for one basic cell, which is composed of a multiplier and an accumulator. The multiplier in the basic cell requires either a large chip area if highmore » speed is desired, or a large amount of time if serial architecture is used. The use of barrel shifters as computational elements in systolic arrays was proposed and studied in detail in this thesis. In the new systolic arrays there are two different structures, parallel and cascaded, that can be used to implement FIR filters. A unique cascaded structure was developed in this study that is shown to have better performance and requires significantly less basic cells.« less
  • This research explores the use of model-based knowledge for real-time production planning and control in a CIM system. A two-level hierarchical control system generates adaptive production trajectories using real-time information. The knowledge-based system monitors the production status, reacts to unplanned events, and performs replanning. The approach is demonstrated over a simulated cell for batch manufacturing with several workstations and material-handling devices. Results show that production goals are met with low variability, and the system adapts well to temporary losses in capacity. Dynamic capacity allocation and dispatching achieve high equipment utilization. An adaptive routing algorithm, using real-time information about the trafficmore » status, achieves low transportation delays. The shell designed for the experiments allows simple integration of heuristic and procedural knowledge for control of dynamic systems, and permits validation of the knowledge base.« less